q7 starting material questions and answers

7/31/2019 Q7 Starting Material Questions and Answers

1/123

Q7A: Questions and Answers

Guidance for industry: Q7A Good Manufacturing Practice Guidance for Active Pharmaceutical Ingredients

Section 1A:Background and History

Q. Based on FDA changing of the meaning of "should," does this mean that Q7A will beapplied differently in the US than in the EU?

No. The change was made by FDA's attorneys to comply with Good Guidance Practices. Therewas no change in meaning intended. The change in the definition of "should," was not intended tohave any impact on how the guidance is applied either domestically or abroad. Remember, Q7Ais a guidance document, but it is not legally binding to either the FDA or the public.

Q. How do you define a "significant structural element" in an API starting material?

A significant structural fragment is that portion of a molecule that contributes to or is responsiblefor the molecule's pharmacological activity. The original definition talked about an importantelement. In Q7A the term "element," was changed to "fragment" to avoid confusion with thechemical understanding of "element."

Q. How close can you get to the API and still call it an API starting material?

The company should document the rationale for the designated API starting material. Where in agiven process the designated regulatory API starting material comes will vary with each process.One of the difficulties was trying to cover all possible situations, because that's nearly impossible.The decision to designate an API starting material is based on more than chemical logic. It is apharmaceutical consideration as well. Remember, there is a patient at the very end of every API

process who is suffering from a disease, and the goal to provide something to get rid of his illnessin a safe and reliable manner.

When you're developing a new chemical process or API process you should be discussing yourchoice of API starting material with the review division. At some point, you should come in andtalk to the reviewers and say, 'this is what I consider to be my API starting material, this is where Iconsider my API process begins." The reviewers are either going to agree with you or disagree,and they might think that it's further back in the process. But, this is something that you shouldagree on early with the FDA review divisions, because it's absolutely critical that you haveagreement with the Agency as to what you define as your API starting material and where yourAPI process begins.

Q. How early should a company discuss with FDA the company strategy, rationale andwhat we consider to be an API starting material for filing?

Q7A does not address filing issues. However, you should meet with the review division as earlyas feasible, to get early input from the reviewers in defining the API starting materials.

Q. If you identify an API starting material vendor, does that vendor need to be audited orqualified?
http://www.pda.org/q7a/q7a/questions/section1a.htmlhttp://www.pda.org/q7a/q7a/questions/section1a.htmlhttp://www.pda.org/q7a/q7a/questions/section1a.htmlhttp://www.pda.org/q7a/q7a/questions/section1a.html


2/123

Section 7.1 states "Manufacturers of intermediates and/or APIs should have a system forevaluating the suppliers of critical materials." The API Starting Material would typically be criticalto the process, so you should have assurance of the quality of that API Starting Material. Forpurchased API Starting Materials, this involves establishing the reliability of the supplier's analysisthrough qualification of the supplier's test results at appropriate intervals, as described in Section7.3. API manufacturers may choose to audit the manufacturer of the API starting material, but thisis not mentioned in nor is it an expectation in Q7A.

Q. Are internal audit reports mentioned in Section 2.4 subject to FDA review during aninspection?

FDA does not generally review internal audits unless just cause exists. FDA has the authority tolook at them, under the statute, if they chose to do so. But they generally do not. Such a requestis rare, and actually, by policy, has to be approved at very high management levels within theFDA.

Q. Concerning the distinction that you made between ICH versus VICH for APIs forveterinary products, which standards do the FDA inspectors use in the plants producingAPIs for US vet products? Are there any differences seen during a PAI, and whichstandards should a plant follow?

Since the VICH at this point has not adopted Q7A, FDA investigators may use Q7A in inspectingmanufacturers of APIs for veterinary drug use. Any issues or deficiency findings will be brought tothe attention of their office of compliance, the Center for Veterinary Medicine. They will then applywhat is commonly referred to as regulatory discretion with respect to those APIs for vet use. Tosummarize, since we lack GMP guidance specifically for APIs for vet use, the best guidanceinvestigators could use is Q7A.

Q. You said that the WHO has now decided to compare the Q7A document with its 1992API GMPs., I heard previously that the WHO was adopting Q7A with no mention of thisevaluation. Is this just a standard procedural issue at WHO or do they have specificconcerns, doubts, problems with Q7A?

The World Health Organization (WHO) was an observer/participant during the development ofQ7A and agreed that it's more scientifically sound, it's updated, and so forth. The WHO also is apublic health agency with a heightened degree of concern for developing countries.The WHO wants to ensure that this guideline is practical for use by regulators in developingcountries where resources are very stretched or even non-existent.

Q. To my knowledge, China, Australia, and India have not had official (observer or

participant) status in previous ICH guidance processes. By participating in Q7A, do theyagree to be bound by it?

China's representation was spotty during EWG deliberations; sometimes an industry person andsometimes a government person. The regulators in China would have to specifically agree oradopt this document.

India did not have regulator representation. There was a conference in Hyderabad that was thefirst conference to discuss the document. Even at the point where Q7A drafts were not supposed


3/123

to be distributed, officials in India had a copy of the Q7A document and it was the subject of theconference. However, the regulators in India have not yet agreed to adopt Q7A.

Q. As a clarification to this question, will the Indian and Chinese regulators be enforcingQ7A on their domestic manufacturers?

The FDA or their European or Japanese counterparts will definitely be using Q7A in China andIndia for materials being imported into the ICH regions. Any intermediate or API imported into anyof the ICH regions would be expected to comply with Q7A, no matter where it originated.However it is not currently expected that Indian or Chinese regulators will be enforcing Q7A onAPIs intended for consumption in India or China.

Q. Since most APIs are manufactured outside the US (80 percent or so), and the FDA'smoving to the MRA for inspections, what training is being given to government inspectorsoutside the US, and what about FDA foreign inspection teams?

We intend to have the same training in Europe as we have outlined here. Nearly all of the

associations, including the European inspectors and the European Commission, who aresupporting or co-sponsoring this event, are participating. It will be a big move forward forEuropean inspectors and for some parts of the industry, in understanding how GMPs are appliedto this aspect of the industry.

Q. Just confirming, the definition of "should" could vary from country to country in thefinal version, for instance the US FDA versus the EU version. What would you recommendfor companies that are international/global? Another person also wanted some practical,down-to-earth words around what "should" should mean. In other words, in here, theperson asks, is "should" in ICH similar to "must," is "should" in FDA similar to "good todo?" In other words, I think there's a little confusion around "should."

At least in the United States, the "should" language in any guidance is a safe harbor. If you follow

it, you should be found in GMP compliance. If you do something different than what is mentionedin Q7A, then you should have information to demonstrate to an investigator that what you'redoing is reasonable. If you're doing what the guidance says, the presumption is you're incompliance and you're doing something reasonable.

The intent of the expert working group was that the definition would be the same in different partsof the world. Both the EU version and the FDA version talk about alternative methods and identifyrecommendations. The language is quite clear in both versions. Q7A identifies GMP expectationsor recommendations for API manufacturing. When we talk about GMP requirements, that'susually referring to a regulation, which is legally binding. So, if something is spelled out in theGMP regulation, 21 CFR 211, that's a requirement, that's legally binding.

Q7A's expert work group provided a lot of examples to help identify what we were looking for. Itwas not meant to be the only way. If we said, "should," our intent was that it's something thatshould be done, but you could do it in a different way if there was another option and you wouldbe able to justify it. The guidance has a lot of flexibility in it to recognize the real world.

Q. What are the main differences between the FDA's March 1998 API draft and Q7A?


4/123

There are several major differences between the two documents. FDA's March '98 draft guidancecovered chemical synthesis manufacturing processes and the later isolation and purification stepsof APIs produced by biotech and fermentation processes. The scope of Q7A is broader.

In addition, Q7A has a chapter on agents, brokers, distributors, and all those other organizations.There's also a chapter in Q7A related to biotech and fermentation. Q7A only covers APIs

intended for human drug products not veterinary drug applications. There were other issues, suchas validation issues, that were either removed or clarified. There were purchasing requirements.There were a number of issues that, quite frankly, made the Q7 negotiation worthwhile for boththe regulators and industry.

Q. The new ICH guideline is a major step forward in ensuring the quality of APIs. How doesthe FDA plan to ensure that they are applied and interpreted consistently by fieldinvestigators, and who, when, and how it will be applied to overseas suppliers?Additionally, it is really encouraging to see the support and participation of FDA in thisimportant series of workshops. How does the Agency plan to carry through thiscommitment within its own organization, for example, training of field staff, key reviewpersonnel in the application of Q7A to API manufacturing?

With Q7A in mind, FDA is planning to update its compliance program for APIs, 7356.002F.However, changes in FDA's inspectional or enforcement policies with respect to APIs are notanticipated.

FDA personnel have attended previous sessions of these agency/industry workshops. In addition,FDA held a one-week session training course for 40 or more FDA investigators in December2001 and is planning a second course in May of 2003. The training includes interpretation of Q7Aand how to conduct API inspections.

Q. FDA plans to revise its September 1991 guide to inspection about pharmaceuticalchemicals and why the need to revise? Does Q7A not replace the 1991 guide?

Since FDA's publication of the Notice of Availability for Q7A, there's been some talk of revisingthe September 1991 BPC inspection guide. However, no decision has been made to proceedwith this revision since some in the FDA still question the rationale for doing this. .

Q. What does Q7A mean?

Q stands for quality, which is one of the four sections within ICH consultation, 7 is the seventhtopic considered under Quality, it's the first GMP topic, but it's the seventh topic considered, andA indicates the first document under the topic of GMP.

Q. USP requires water used in the manufacture of parenterals to be water for injection,

WFI. Does Q7A conflict with this?

No. If you go back to the scope of Q7A, it clearly states that the sterilization and asepticprocessing of sterile APIs are not covered by this guidance

Q. Now that we have Quality System Inspections and Q7A, can you describe what youenvision as a typical FDA inspection of an API plant?


5/123

The concepts embodied in Q7A grew from concepts used for years. It will not substantiallychange FDA's audits and inspections of API manufacturers. It does establish clearer guidelines toassist all parties.

Q. Any background, history on why the accountability for lab records was not included in

Q7A? For example, numbered pages in a lab bound notebook or sequential lab sheets thatcannot be duplicated.

The accountability of lab records stems from the Barr decision and the falsification of laboratorydata. Q7A is an international document; it is not just a U.S. document. It provides good, soundguidance on documentation practices, but it does not specify sequential numbered sheets or thelevel of documentation that has come to be used in the U.S. for lab records since the Barrdecision. Q7A allows flexibility in meeting documentation practices.

Q If the raw material is commercially available, and if this raw material is the API only afterpurification, so we're basically one step removed, does the raw material manufacturer fallunder Q7A?

It sounds like you're bringing an API into your facility and just further purifying it. Does Q7Aapply? Yes, because, basically, you're bringing in the active ingredient in an unpurified form andsubjecting it to the purification. So, the manufacturer of the crude API would probably fall underQ7A.

Section 1B:Introduction

Q. If the Act does not make a distinction between API and dosage drug product and 21CFR 211 is a regulation and ICH Q7A is a guidance document, then an FDA investigatormay still choose to inspect an API manufacturer using specifically and only 21 CFR 211.

This is part of the reason why we have FDA investigators present in the audience at this trainingsession and we've had them present in the other training sessions of Q7A. It's also part of thereason why ORA and CDER, are sponsoring FDA national training courses to discuss Q7A, andtelling investigators and industry the same thing when auditing or inspecting API manufacturers.

However, if you find yourself in this situation pull out the preamble to the 1978 GMP regulationsand show the investigator the commissioner's response to comment 270 where it clearly saysthat 211 does not apply to bulk chemical manufacturing.

Q. Did the Expert Working Group discuss a definition of timely? If so, please comment.

We considered the Barr decision definition of "timely" in terms of closing out deviations within 30days. If it is not possible to do it within 30 days, then it would be wise to write an interim report. It

depends on the issue and the urgency or the health implications of what's timely. If it's a situationwith potential health implications for patients down the line, then "timely" is certainly faster than30 days. If the situation involves a complicated corrective action, maybe it'll take more than 30days. But you need to think about it in those terms.

Q. If equipment is closed, do you usually need environmental controls?

If the equipment is contained or totally closed whereby the API or intermediate does not see theenvironment, generally there is no rationale for environmental controls.
http://www.pda.org/q7a/q7a/questions/section1b.htmlhttp://www.pda.org/q7a/q7a/questions/section1b.htmlhttp://www.pda.org/q7a/q7a/questions/section1b.htmlhttp://www.pda.org/q7a/q7a/questions/section1b.html


6/123

Q. When and where do we start applying CGMPs? For example, we may use isopropanolto make the reaction of our API or to purify API. We do not manufacture isopropanol, sodoes the manufacturer of isopropanol have to follow all the CGMPs that apply to us? Doesthe manufacturer of any excipient have to follow the CGMPs as we do?

This is one of those examples where using this material, isopropanol, based on its intended use,

could be classified as a solvent or an active pharmaceutical ingredient. If the company only usesisopropanol in reactions to purify the API, then clearly, the isopropanol would be categorized as asolvent or a raw material. However, if you are a drug manufacturer and you are receivingisopropanol, and you're producing rubbing alcohol, which is classified as a drug, then under thatscenario the isopropanol would be the active pharmaceutical ingredient. That's why we have "theintended use" clause in the Q7A definition because certain materials, depending on their intendeduse, could or could not be classified as API.

Q. Does Q7A apply to excipients manufacturer?

No, Q7A only applies to active pharmaceutical ingredients as defined in the guidance document..However, there is other guidance available, such as the voluntary guidance developed by theInternational Pharmaceutical Excipients Council (IPEC). The inclusion of language dealing with

excipients was discussed by the Q7A Expert Working Group, but it was decided to focus on APIs.

Q. How are the FDA investigators actually receiving your advice and what are we doing toprovide specific training?

ORA has sponsored some national API training courses for field investigators in addition toallowing agency personnel to attend these workshops.

Q. I've got a question here about critical deviation and non-conformity or non-conformance and why were these not harmonized in the document?

Members of the expert working group thought that it was clearer to use deviations or non-conformances rather than using just one of those terms.

Q. When can we expect the industrial guidelines that went into the development of Q7A tobe integrated into the local regulatory agencies. For example, as part of the CFRrequirements that would insure compliance.

I assume the person is talking about the various documents mentioned that were used to developQ7A, the PhRMA document, the FDA's March '98 draft, the PICS document. All these documentshave been superseded by Q7A. What's being integrated into the regulatory mechanisms in thethree ICH regions is Q7A.

Q. The FDA version of ICH Q7A. Why did the FDA version format the section numbers orleave out the section numbers?

That's basically because of the regulation on Good Guidance Practices (GGPs) which requirethat FDA guidance documents follow a particular format.

Q. If GMP does not start until the starting material is used in the manufacture of APIs, whythen are starting material manufacturers inspected for GMPs?

Starting material manufacturers are not routinely inspected by the FDA.


7/123

Q. Are there any guidance requirements within Q7A for starting materials?

Section 1.3 defines API starting materials and states "This GMP guidance does not apply to stepsprior to the introduction of the defined API starting material."

Q. Radiopharmaceuticals. Are APIs involved in radiopharmaceutical production? Areradiopharmaceuticals excluded from the scope of Q7A because of their unusual process?

Radiopharmaceuticals are excluded from Q7A because of the uniqueness of those processes.

Q. Can an investigator issue observations based on the Q7A guidance?

All the guidance documents, not just Q7A, clearly state that a guidance document is not legallybinding in the way that a regulation is. A binding regulation means that you must comply with it orelse you are in violation of the regulation, whereas a guidance document, like Q7A, providesrecommendations. You could employ alternatives as long as you show these are equivalent.

Q. For a company involved in production of both APIs and drug products, how should oneuse Q7A along with other guidance?

For the APIs, Q7A should be used. For drug products, 21 CFR 211 should be used.

Q. Would the acid used to convert a free base into a salt form of a drug substance beconsidered an API starting material? Significant structural fragment?

No. The acid would be considered a raw material but would typically not be considered an APIstarting material. Regarding the significant structural fragment, it would depend on the specificacid used, but again typically an acid used only for conversion of a base to salt would notcontribute the significant structural fragment to the API.

Section 2:Quality Management

Section 2.1

Q. What is your opinion regarding API manufactured with limited resources in which thequality unit if comprised of only one person?

Section 2.1 states that "there should be a quality unit that is independent of production and thatfulfills both quality assurance (QA) and quality control (QC) responsibilities. The quality unit canbe in the form of separate QA or QC units or a single individual or group, depending on the sizeand structure of the organization." So for a very small organization it might be reasonable to havea single person having full responsibility for quality.

Q. "Independent from production." Does having different managers constituteindependence? If not, at what point in an organization is it acceptable for quality andmanufacturing to report to the same person?

The fundamental idea is not to have conflict of interest. For example, the senior vice president forquality should have the independent authority to approve or require changes.
http://www.pda.org/q7a/q7a/questions/section2.htmlhttp://www.pda.org/q7a/q7a/questions/section2.htmlhttp://www.pda.org/q7a/q7a/questions/section2.htmlhttp://www.pda.org/q7a/q7a/questions/section2.html


8/123

Q. Is there an outside agency that will certify GMP compliance much like ISO 9000 series?

No.

Q. Should quality units actually develop definitions and examples of deviations and criticaldeviations?

It is important for each manufacturer to define what is critical to each API process. If you deviatefrom something that you have defined as a critical parameter, then that becomes a criticaldeviation.

Unfortunately, as companies continue to run a process over the years, they learn about theprocess and uncover new factors or parameters that they had not originally identified as critical,that indeed affect API quality. So it's important to continue to update your list of criticalparameters with knowledge that you acquire over the years. That's part of the purpose of theProduct Quality Review discussed in 2.5.

Section 2.2

Q. Why should rejection of an API not be delegated? In other words, does manufacturinghave the right to reject if something goes out of specification?

Section 2.2 (1.) stipulates that the quality unit is responsible for releasing or rejecting all APIs andintermediates for use outside of the manufacturing company and that these responsibil itiesshould not be delegated. The reason for this is that these decisions by the Q.C. unit should not bequestioned or overridden by production or other departments.

Although not clearly stated, this section implies that the QC unit can delegate the release orrejection of intermediates used internally by further manufacturing.

Q. Release authority for intermediates can be delegated to production except forintermediates that are sold, what about intermediates that are shipped within companieswithin the same corporation?

Section 10.2 reads, "APIs and intermediates can be transferred under quarantine to another unitunder the company's control when authorized by the quality unit (s), and if appropriate controlsand documentation are in place." So as long as it's still in the same corporation or asubcontractor, its foreseeable that production could release intermediates provided this isauthorized by the quality unit.

Q. If the laboratory is part of the quality unit, must the procedures, test methods,equipment, etc., be approved by the quality unit in addition to the person in thelaboratory?

Section 2.2 states that the quality unit is responsible for "approving all specifications and masterproduction instructions, approving all procedures affecting the quality of intermediates or APIs",and "approving changes that potentially affect intermediate or API quality." Q7A simply defines


9/123

activities that should be performed by the quality unit. Q7A does not specify who within thatquality unit should perform specific activities. So if your quality control laboratory is part of thequality unit, there is no expectation that there be an additional person inside the laboratory whoalso approves laboratory operations.

Q. In the non-delegateable activities, does that mean that I cannot have a quality controllaboratory as an external supplier?

No, that's not what it means. Testing is not listed as one of the responsibilities that should not bedelegated. Section 2.2 (1) clearly states that the quality unit should not delegate the release orrejection of APIs. So regardless of who is conducting the testing, it's up to the quality unit toreview the results and to make the release.

Q. Do the responsibilities not to be delegated by QA apply to both clinical and commercialAPIs? For example, do manufacturing instructions require QA approval prior tocommencement of the manufacturing step?

Responsibilities not to be delegated apply to both except where Section 19 indicates otherwise.Manufacturing instructions is one of the areas where expectations are specifically different forAPIs intended for use in clinical trials. Section 19 states that the expectation for documentation isdifferent for clinical materials. It does not call for a Batch Production Record. Section 19.5 states,"Production can be documented in laboratory notebooks, batch records, or by other appropriatemeans." It is the role of the quality unit to ensure that appropriate documentation is kept.

Q. What is the role of the quality unit in approving changes to master productioninstructions during clinical trials? What is the quality unit's role in reviewing andapproving the master production instructions in phase I and phase II?

In section 19.7 under clinical trials, Q7A states: "Changes are expected during development, asknowledge is gained and the production is scaled up. Every change in the production,

specifications or test procedures should be adequately recorded." So the role of the quality unit isto make sure that there is a system for recording all changes and tracking when, in the process,those changes were made.

Q. What should approving of contract manufacturers involve?

A contractor, regardless of the activities they perform, is an extension of the manufacturer'soperations. Therefore, the manufacturer is still responsible for ensuring GMP compliance forthose activities that have been contracted out. Again, just the fact that you contracted an activitydoes not mean that you no longer have responsibility for that activity. Refer to Section 2.2 (8).

Q You mention the audit expectation for API starting material suppliers. What do youexpect from audits of API suppliers by drug product manufacturers?

With Q7A we now have a GMP guidance document that can be used when drug manufacturersaudit an API supplier or manufacturer. It's the same document that the API supplier uses, and it'sthe same document that the regulators use. For the first time in history, we have a GMPdocument, which both industry and regulators can use, and everybody should have a clearunderstanding of the GMP expectations.


10/123

Q. With regards to quality and production responsibilities, it is understood that the qualityunit cannot delegate all of its responsibilities, but can production delegate itsdocumentation review of all documents to the quality unit?

No.

Q. Equipment and computerized systems are not part of the term material, so suppliers ofcritical equipment and computerized systems do not need to be approved.

That is an incorrect interpretation of Q7A. Section 2.2 (9) states that the quality unit has theresponsibility to approve any changes that could affect the intermediate or API. So if the newequipment or new computerized systems could potentially affect the API in any way it should beapproved by the QC unit.

Q. Typically an API manufacturer performs quality tests on intermediates. If it passes, thenQC will release the intermediate for the next synthesis step. The quality unit then reviews

the entire batch records at the end and releases the final API. It seems that Q7A requiresthat the quality unit review the batch records and test results at each step, prior to movingto the next step. This may not be practical in some cases.

That's an incorrect interpretation of Q7A. Section 6.7, the final paragraph, states that "productionand laboratory control records of noncritical process steps can be reviewed by qualifiedproduction personnel or other units following procedures approved by the quality unit (s)."

In addition, Section 2.2 (3) states that the quality unit is responsible for "reviewing completedbatch production and laboratory control records of critical process steps" before the release of theAPI for distribution. It is up to the company to define which steps are critical. So the individualrelease of the intermediates can be performed by production according to systems that havebeen approved by the quality unit, and the timing for the quality unit review of the batch records is

based on the release of the API.

Q. Am I correct though, in terms of that release of intermediates, there is also arequirement that the quality unit review any deviations, so the production group could bereleasing intermediates, but only if the intermediate is good, and they need theinvolvement of the quality unit for deviations?

Yes, the quality unit should review the deviations.

Q. Is there no need for environmental monitoring such as occupational hygienemonitoring for API manufacturing?

Q7A does not address safety considerations for manufacturing either from an OSHA(Occupational Safety and Health Administration) point of view or from an industrial hygiene pointof view. So any monitoring that needs to be done to be in compliance with OSHA is beyond thescope of Q7A.

Q. Please further clarify the scope of QA and the approval of master productioninstructions versus the production review and approval.


11/123

Production is responsible for writing and reviewing the master production instructions and forassuring that these instructions contains all the details that production feels it should contain. TheQuality Unit is responsible for final approval of master production instructions. .

Q. Q7A says the quality unit should review and approve all appropriate quality relateddocuments. How much scope did the expert group give to this statement? It's easy to see

why production and laboratory documents, but what about maintenance documents?Should the quality unit review and approve a procedure that, for example, tells a mechanichow to put oil into a pump? Another example, should they approve housekeepingprocedures?

What you really need to do is balance good judgment and common sense. Typically the qualityunit should review the procedures having to do with scheduling of preventive maintenance andhow the records on these preventative maintenances are kept. Then the quality unit wouldtypically perform a check of some sort to ensure that preventive maintenance is done on time. Butthe quality unit does not necessarily need to review the detailed procedure on how thatpreventative maintenance is done. The same thing for housekeeping, where the quality unitshould review the procedure defining how often, who's going to do it, that sort of thing, but thedetails of how it gets done does not necessarily need to be reviewed by the quality unit. The one

exception to that would be if you've got micro specifications and the housekeeping really doesimpact your API .

Section 2.5

Q. Is it ever appropriate for the quality unit to make sure that the Product Quality Review isperformed by another group or is this an activity not to be delegated. For example, reviewof in process controls, critical API results performed by protocol, approved by the qualityunit, issued from technical operations, a group that does not report directly to the qualityunit.

Q7A Section 2.2 (15) states that the quality unit's responsibility for "performing product qualityreviews" should not be delegated. This wording does not imply that the quality group perform allthe work. Other groups may pull together the information and data, but the quality unit should atleast review that data and make the final conclusions.

Q: What is expected with regard to trending deviations?

Typically that would be done as part of your product quality review. Part of the reason for trendingdeviations is to look for problems in your processes. Deviations, even the non-critical deviationscan be a good indication of the need to make some equipment changes or some sort ofmechanical changes so that the operator can follow the instructions without deviating. Thedeviations can also be a good indicator of when training needs to be done. Q7A is not explicitabout trending of non-critical deviations, but it is a good business practice to do so.

Q. Would it be acceptable to use annual product review as an opportunity to determinewhich processes and products require revalidation?

Certainly. In fact, this is addressed in Q7A Section 2.5 under product quality reviews: "The resultsof this review should be evaluated and an assessment made of whether corrective action or anyrevalidation should be undertaken." So clearly it is intended that data from the annual productreview could be used to determine if revalidation is necessary.


12/123

Q. Who is responsible for making recommendations for corrective actions for revalidationas a result of yearly product quality review?

That would typically be a joint decision between the quality unit, technical services or engineeringservices or wherever your technical expertise resides and production.

Q. Under Section 2.5, the product quality review, what is the expectation for the review ofcritical in process control and critical API test results?

What's expected is that you look at the trends. Are impurities increasing or decreasing? Are thetest results all over the place? Simply a review of the trend is what is typically expected there.This can be done with a table of data or a graphical plot depending on the amount of data to bereviewed.

Q. Similarly, what is expected in the product quality review for the review of the accuracyof the corrective actions?

What's expected is that you look to see, did the correction that you put in place work? Did theproblem recur again or was that correction action effective? The other thing you should look for iswhether putting that corrective action in place causes other problems.

Section 3:Personnel

Q. In the past there has always been a difference between the qualified person as definedin Europe and his or her responsibility and the responsibilities deriving from the educationof a person are being seen in the U.S. Can you comment whether this continues to exist

and is being seen differently? Basically we're looking at the European question of aqualified person versus here.

Q7A does not address the issue of qualified person. Section 3.1 reads, "There should be anadequate number of personnel qualified by appropriate education, training, and/or experience toperform and supervise the manufacture of intermediates and APIs." In the European Union thereis a requirement that the final release of medicinal products (finished pharmaceuticals) be doneby a Qualified Person. This Qualified Person is typically a pharmacist. The EU does not require aQualified Person to release APIs.

In Japan there is a requirement that a security pharmacist has to release all the APIs.

Q7A can be overridden by regional requirements, but the qualified person was intentionallyexcluded in Q7A. It is important that person have the appropriate education and experience forthe processes and products that are being reviewed and released. In some cases it makes moresense to have a chemistry or biochemistry degree than a pharmacy degree.

Q. Does the Q7A Section 3 on personnel directly apply to laboratory and QA personnel,i.e., adequate number of qualifications, etc., or is it intended to apply only to themanufacturing personnel?


13/123

. Manufacturing in the U.S. has come to be a synonym with production. Q7A goes back to the useof "manufacturing" in good manufacturing practices. GMPs don't just apply to production; GMPsapply to all aspects of making the API. "Adequate personnel" applies to manufacturing becausethat's everybody. That's your lab people, warehouse people, everybody that has to do with thecontrol, preparation, production, testing and distribution of the APIs, and it also applies tolaboratory.

Q. What personnel are covered by the training requirement?

Training applies to any personnel who are engaged in manufacturing APIs. Remember thedefinition of manufacturing includes anyone involved in the receipt of materials, production,packaging, repackaging, labeling, relabeling, quality control, release, storage and distribution ofAPIs and related controls. It also applies to your maintenance mechanics, anyone who is workingwith any equipment that affects the API. This also applies to consultants and temporaryemployees.

Q. Is training using the Internet acceptable? If so, how could it be documented?

You can utilize any form of training that you find to be meaningful. Documentation should include

what the training covered, who provided it, how the training was given and to whom.

Q. Does training record maintenance differ significantly from what is specified by 211?

If you compare the language in 211.25 and 211.5 with what's in Q7A, there are no significantdifferences in terms of the expectations for training. There is more detail in Q7A than in CFR 211.

Q. Can training records for Q7A covered operations be covered electronically withoutbeing Part 11 compliant?

Electronic training records should meet the expectations for computerized systems in Q7A,Section 5.4.

Q. Please discuss ways in which investigators will assess whether there are an adequatenumber of personnel.

When deviations occur, an investigator may question whether enough people are involved in thework or when investigations aren't being completed in a timely manner whether there are enoughpeople conducting the investigations. That's something that each company should look at on aperiodic basis. A good indicator may be the time required to complete investigations andimplement corrective actions.

Q. Are regulators looking for disciplinary action against basic GMP errors made by trainedand qualified operators?

Q7A does not address disciplinary action against people who make mistakes, nor does it prohibityou from taking such action if it's your company's policy.

You should be very careful of overusing "employee error" or "operator training" as a reason forfailure in lieu of conducting a real investigation. It's not always the person that's causing the error.Sometimes the equipment is set up in a manner that the person can't do it any other way.Investigators will raise questions about determining the real root cause if you have too many


14/123

deviations where the problem is always seen as operator error. Regulators are looking for rootcause identification and true corrective action, not disciplinary action.

Q. On training, what is your interpretation of a person qualified to conduct GMP training?

As per Section 3.1, a person conducting GMP training could be qualified by appropriate

education, training and/or experience in the subject matter being taught.

Q. Does the fact that training should be periodically assessed mean testing of individualsor watching the individual perform the operation?

It could be done either way. Q7A tells you what should be done, but it isn't specific on how itshould be done. One suggestion is to use your product quality review to look for any trends. Arethere a large number of errors on a particular operation? Do you have an indication that additionaltraining needs to be conducted? There are a number of ways that you can look for gaps in yourtraining and repeated testing is not always the most effective way to address that. Also a reviewof your deviations can be used as an indication about the adequacy of your training program. Ifyou're continually writing up deviations and you attribute these to poor operator training, then your

training program might need to be improved.

Q. Please expand on the definition of consultant. Can we exclude persons who arecontractors that work under direct supervision of company employees?

That description sounds like temporary employees, and you're not excluding them, you're just notcalling them consultants. Temporary employees would be covered under Q7A Section 3.1, whichreads, "The responsibilities of all personnel engaged in manufacture of intermediates and APIsshould be specified in writing. Training should be regularly conducted by qualified individuals andshould cover, at a minimum, the particular operations that the employee performs and GMPs as itrelates to the employee's functions. Records of training should be maintained."

Q. Please give an example of what qualified means to a laboratory technician. Is it tenyears experience? Is it a BS in chemistry? Is it documented training on certain LC and GCanalysis or is it some sort of combination of all of that?

As per Section 3.1, a laboratory technician can be qualified by appropriate education, training,and/or experience (e.g. on the job training.) It depends on what that laboratory technician is doingfor you. It's up to you to define what you need for the different tasks that are being performed. It'salso important that these people keep their training and their education up to date.

Q. Please clarify the meaning of "training should be periodically assessed." Does thismean competency assessment of those being trained, i.e. post-training assessment?

The EWG spent a lot of time talking about what we meant with this language in Section 3.1.Some members of the EWG felt very strongly that you can assess training while others felt thatthis was not feasible.

One way to assess training is to look at the accuracy with which the employee is performinghis/her job. If the employee is making a lot of mistakes or deviations you may need to provideadditional training. Some companies use a post-training testing program. There are multiple waysyou could assess the adequacy of the training. The objective is to show that the training you areproviding is actually being received and implemented.


15/123

Section 4:Buildings and Facilities

Q. In a multi-step API manufacturing process, can charging operations for earlier non-critical steps be done outdoors, as long as the last several critical steps are done indoors?

Under Section 4.1 of Q7A, it is very clear that where the equipment itself (e.g., closed orcontained systems) provides adequate protection of the material, such equipment can be locatedoutdoors. That is not restrictive at all as to steps. It depends on whether the equipment providesadequate protection of the material.

Q What is the expected frequency of testing of High Purity Water systems, such asPurified Water with Endotoxin Control: daily, weekly, bi-weekly, monthly, etc., for point ofuse sampling?

Q7A does not specifically address frequency of testing of high purity water systems. Thefrequency of testing for both endotoxin and microbial testing of water systems will vary from onecompany to another. A company should assess what is most practical for their particularoperation. The frequency of testing for both microbial and endotoxin attributes will depend uponthe design of the water system, the quality of the source water and other factors

Q. For Intermediate processing, comment was made that the processing of potentiallytoxic, i.e. pesticides, was not that significant an issue. Can a company support such anactivity?

Yes. Some APIs may actually be derived from an intermediate that is a pesticide or an herbicide.This is possible at an intermediate stage. (Section 4.4)

Q. Is Purified Water or Water for Injection the quality of water you are requiring orrecommending for use in API production?

As per Section 4.3, "water used in the manufacture of APIs should be demonstrated to besuitable for its intended purpose." So, it is up to the manufacturer to determine what quality ofwater should be used at various stages of the API process and establish appropriatespecifications. Potable water is the minimum quality of water to be used in API manufacturing.(Section 4.3)

Q. Permanently installed pipe work identification: would documentation be a P&ID, orwould a process flow diagram be an adequate substitute for labeling individual lines?

It depends. Q7A Section 4.2 states, "permanently installed pipe work should be appropriatelyidentified." It does not address how this should be accomplished. It's up to the manufacturer todetermine this. The identification should typically be visible to the persons (operators) working inthat area.

Q. Which engineering steps during the planning of a new API plant should be kept,documented, and verified?

Q7A does not get into that level of detail. Sound engineering and scientific practices shoulddictate what data should be retained.


16/123

Q Does the manufacture of API's require the establishment of air quality standards, HVACsystems specifications to insure no dust migration during the charging of componentsused in API manufacturing from one reactor to another?

Section 4.2 of the guidance states that "adequate ventilation, air filtration and exhaust systemsshould be provided, where appropriate. These systems should be designed and constructed to

minimize risks of contamination and cross-contamination and should include equipment forcontrol of air pressure, microorganisms (if appropriate), dust, humidity, and temperature, asappropriate to the stage of manufacture. Particular attention should be given to areas where APIsare exposed to the environment."

If the API process is essentially a closed-system, this would minimize the need for establishmentof air quality standards and specific area classifications.

Q. Drug product manufacturers rinse manufacturing equipment with purified water. ShouldAPI manufacturers rinse finishing equipment such as mills and blenders with purifiedwater?

Q7A does not specify the quality of water that should be used in any process or cleaning step.Once you determine what quality of water is suitable for a particular step, you might want to usethe same quality of water to rinse the equipment. (Section 4.3 & 5.2)

Q. Is microbiological control in the environmental clean room necessary if the room is notused for injectable grade products?

Section 4.2 states that if microbiological control is necessary, your facilities should be designed tominimize the possibility of microbiological contamination. Sterile APIs are outside the scope ofQ7A.

Q. The new EMEA note for guidance on water has some clauses around APIs and waterqualities that are tighter than the Q7A guidance. How should companies interpret thesetwo guidances?

The Q7A expert working group had extensive discussions over a number of meetings around theworld about the quality of water, and the final wording recognizes that in many circumstancesthere was not a good justification to expect more than potable water quality.

We considered a lot of issues and opinions when developing this language, but the bottom linehere is whenever you're asking about the quality of water, you always have to ask what is thewater being used for? If you're using potable water at a very early process step, its impact maybe minimal, even if you're using it at a later processing step, its impact may be minimal. As per

Section 4.3, the manufacturer basically needs to determine that the quality of water they're usingat any particular step is suitable for its intended purpose.

Q. Regarding closed systems for APIs in the same room allowing for multiple products tobe manufactured in that same room, is that just different lots of the same product ordifferent chemical entities or reactions?

It is uncertain what the question means by the "same room" in an API facility.


17/123

Section 4.1 states that facilities should be designed to minimize potential contamination and crosscontamination. That's the intent. There is no reason why you could not have different processesor different intermediates or APIs processed side-by-side in closed reactors.

Obviously, a packaging operation is a different scenario. Having side-by-side packagingoperations without some sort of a barrier or containment between the lines is not an acceptable

GMP practice. The intent is to minimize cross contamination from one process to another.

Q. We've heard Europe has defined a fourth level of water purity called highly purifiedwater, which seems to be aimed toward the production of APIs. What impact will this haveon the other ICH regions and Q7A if any?

Q7A, like any GMP document, does not define standards. API manufacturers should establishadequate specifications for process water, but Q7A does not define or specify the use of purifiedwater, highly purified water, or WFI. This is outside the scope of Q7A.

Q. Does the manufacturing facility need to be qualified to be at least class 100,000 for APImanufacturing?

No. See Section 4.2.

Q. Section 4.4. For material of high pharmacological activity or toxicity, dedicatedproduction areas should be considered unless there is validated inactivation or cleaningvalidation. Is there an amount or a method for determining if your compound is consideredhighly potent?

Q7A does not define potent or a potent compound. There are some definitions in the OSHArequirements in terms of the various containment areas. (Section 4.4)

Q. Would validation of critical HVAC controls be expected if there were a separatevalidated alarm and monitoring system?

As per Section 4.2, all utilities that could affect product quality, to include heat, ventilation, and airconditioning systems, should be qualified and appropriately monitored.

Q. In design and construction, is there any topic under the subtitles regarding localhazards because of adjacent units, adjacent companies, which are involved in non-pharmaceutical operations? Say dyes for the textile industries or rubber industries and thelike that are located in the same chemically zoned industrial district as the plant?

No. Q7A does not specifically address this issue. However, Section 4.1 states that buildings and

facilities used to manufacture intermediates and APIs should be located, designed andconstructed to minimize contamination or cross contamination. So if you're in an area in whichfor, one reason or another, you believe there is a possibility of contamination from a neighboringchemical plant, then you have the responsibility for ensuring that the facility, equipment and theprocess is protected from potential contamination.

Q. If you're using electronic systems for control of materials, exactly what physicalsegregation or physical separation of products and areas is needed?


18/123

Section 4.1 states there should be defined areas or other control systems for a variety ofactivities. If you have an adequate documentation or electronic system that provides appropriatecontrol and handling of materials, there's not an expectation for physical separation of thematerials. In addition, Section 7.2 states, "A system should be in place to identify the status ofeach batch." This can be a documentation or computer system.

Q. For nitrogen gas, which is being used for vessel-to-vessel transfer. What qualification isrequired? Anything besides IQ, OQ and PQ?

Section 4.2 states, "All utilities that could affect product quality should be qualified andappropriately monitored, and action should be taken when limits are exceeded." So in addition toqualifying a system, there should be some sort of monitoring.

Q. What does adequate protection in Section 4.1 and adequate filtration in Section 4.2mean?

Q7A does not address this because it's up to the manufacturer to determine what is adequateprotection for their particular processes. Section 4.1 states, "Where the equipment itself, (e.g.,

closed or contained systems) provides adequate protection of the material, such equipment canbe located outdoors." If it's a reactor, and it's completely closed, that's probably fine. If it'sequipment that you are opening outside to sample or to introduce materials, you may need sometemporary protection where the system is being opened.

Q. There are several questions here that relate to standards for clean room classes andare asking about regulations or specific guidances that identify either where they'reneeded or about design requirements for room classifications, Class 10,000, etc.

Q7A does not define standards for clean room classes nor specify classes for API processes orsteps. Within the United States, no regulation or guidance has been issued defining the types ofclean rooms or room classifications for APIs.

Section 5:Process Equipment and CleaningQ. Wherever possible, food grade lubricants and oils should be used. Does that mean it'sokay to get into the batch?

The reason that was put in to Q7A is that you should be using materials that will cause the leastamount of harm if a small amount gets into your batch. If, however, you have catastrophic sealfailure and a large amount of lubricant gets into the batch, you really should investigate this as adeviation and determine its impact on the quality of the API or intermediate. Yes, that does meanthat small amounts may get into your batch, but it is not intended to imply that you don'tinvestigate failures or process deviations

Q. What do you recommend in terms of status indicators for production equipment? Giveme examples.

As per Section 5.2, "Equipment should be identified as to its contents and its cleanliness statusby appropriate means." This can be accomplished by physically labeling the equipment or bycomputer controls.

Q. Regarding equipment identification. Is it considered necessary to identify transfer lineswith direction of flow?


19/123

It might be an issue for safety, but it is not addressed in Q7A.

Q. What was EWG's thinking along the lines of consumable items such as inline filters, O -rings and gaskets? Is there an expectation that they are treated as job aids or equipment?

It would be reasonable to regard them as part of the equipment and include these in the

equipment change control procedure. If you exchange these items and they have a differentcomposition, different material, they can have a significant impact on the intermediate or API.

Q. Is the performance of preventive maintenance required for each piece of GMPequipment?

Any piece of equipment is subject to wear and tear and a company should have a preventivemaintenance program for all pieces of equipment. It doesn't make a lot of sense to buy a veryexpensive piece of equipment and not have some type of scheduled maintenance of thatequipment. Again, this is one area where good common sense practices dictate that this is agood thing to do. It will minimize sporadic or catastrophic failures of equipment, and it willmaximize the use of that equipment and its shelf life.

Q. Under preventive maintenance, is it acceptable to let certain equipment to go to failure,such as piping, manual valves, etc?

As per Section 5.2, "Schedules and procedures (including assignment of responsibility) should beestablished for the preventative maintenance of equipment." If manufacturers allowed equipmentto run until the equipment broke down without concern for maintenance, the companymanagement would bring it to your attention well before the regulators would. Some regulatorswould conclude that periodic or frequent equipment failures due to lack of a preventivemaintenance program is a GMP deficiency.

Q. Are flexible hoses that are repeatedly used considered major equipment, thereforerequiring maintenance?

Flexible hoses are not major equipment, but they should be adequately identified, maintained andcleaned. Section 5.2 emphasizes that "written procedures should be established for cleaningequipment and its subsequent release for use in manufacture of intermediates and APIs."

Q. Does Q7A provide for the matrix approach to equipment cleaning validation, or arecompanies expected to perform a separate cleaning validation for each compound?

Yes. Section 12.7 states: "Validation of cleaning procedures should reflect actual equipmentusage patterns. If various APIs or intermediates are manufactured in the same equipment and thesame process cleans the equipment, a representative intermediate or API can be selected forcleaning validation. This selection should be based on the solubility and difficulty of cleaning and

the calculation of residue limits based on potency, toxicity, and stability."

Q. Is it expected to establish a maximum time elapsed from the end of processing tocleaning? Give some examples where this is not appropriate.

According to Section 5.2, procedures should include "establishing the maximum time that mayelapse between the completion of processing and equipment cleaning, when appropriate." Eachcompany should evaluate the need for establishing maximum elapsed time based on the difficultyof cleaning the equipment after prolonged standing. The GMP issue is to assure that your


20/123

cleaning procedure is capable of removing residues that remained in the equipment for anextended period of time after its use.

Q. Is there an expectation to establish time limits for clean equipment before thatequipment needs to be recleaned?

Once the equipment is cleaned, it is important that you protect the equipment from contamination.Section 5.21 states that there should be an "inspection of equipment for cleanliness immediatelybefore use, if practical." There is no expectation for a time limit after cleaning.

Q. Cleaning methods - are you looking for the active ingredient, the raw material, and by-product in the cleaning methods?

Generally in cleaning validation you are attempting to remove residues of the compoundpreviously manufactured in the equipment and cleaning agents, if these are used.

The objective is to look for compound residuals that shouldn't be in the next intermediate or API.

Anything that you don't expect to be there is what you're trying to f ind and remove from theequipment. Obviously, if you don't use cleaning agents, then testing for cleaning agents would notbe an issue. However, if a company uses a solvent as a final rinse for a piece of equipment, andthat same solvent is used in the next process step, they would not be expected to test forresiduals of this solvent.

Q. Does the ICH provide guidance on the determination of cleaning limits, residual limits?

Calculation of cleaning limits is not addressed in Q7A. Section 12.74 emphasizes that residuelimits should be "practical, achievable, verifiable and based on the most deleterious residue." Thelimits can be established "based on the minimum know pharmacological, toxicological, orphysiological activity of the API or its most deleterious component."

Q. What is the group's opinion of detergent cleaning versus solvent cleaning?

This issue is not addressed in Q7A, although it was discussed by the EWG. If the solvent used toclean equipment is the same solvent that will be used in the next manufacturing step, it would notbe of concern. If you are using a detergent for cleaning equipment or a solvent that is not used inthe next step, cleaning methods should demonstrate removal of residuals from the detergent orthe solvent.

Q. Solvents used for cleaning equipment. May these be recovered?

According to Section 14.4, "solvents can be recovered and reused in the same process or indifferent processes, provided that the recovery procedures are controlled and monitored toensure that solvents meet appropriate standards before reuse." This includes recovery ofsolvents used for cleaning. Of course, you should establish appropriate test specifications andquality criteria for the recovered solvents.

Q. Visual and analytical verification after cleaning, is one exclusive to the other?


21/123

Section 12.7 states that "equipment cleanliness can be monitored by analytical testing and visualexamination, where feasible." Visual examination can allow detection of gross contaminationconcentrated in small areas that could otherwise go undetected by sampling and/or analysis."Generally, some combination of analytical testing and visual exams are employed.

Q. For a non-sterile API intended for use in a sterile drug product, what level of equipmentsanitization and/or sterilization is expected and how verified?

This issue is specifically addressed by Section 12.7 which states that "equipmentcleaning/sanitation studies should address microbiological and endotoxin contamination for thoseprocesses where there is a need to reduce total microbiological count or endotoxins in the API, orother processes where such contamination could be of concern."

The primary issues with a non-sterile API intended for use sterile drug products are the possibilityof microbial and endotoxin contamination of the API and subsequent incorporation into the drugproduct. Endotoxins, once in the API, are generally not removed by drug manufacturingprocesses.

If you're producing an API that is susceptible to microbial contamination or supports microbialgrowth, it may be prudent to sanitize or sterilize equipment to minimize microbial and endotoxinlevels.

Q. Is cleaning validation expected to take place when the first batches are produced forcommercial use (process validation), even when such batches are produced on a pilot orsmall scale?

Section 19.3 emphasizes that during all phases of clinical development, including the use ofsmall-scale facilities or laboratories to manufacture batches of APIs for use in clinical trials,procedures should be in place to ensure that equipment is calibrated, clean, and suitable for its

intended use." Due to the nature of the manufacturing equipment used (e.g., laboratoryglassware, small scale equipment or pilots scale equipment that is easily disassembled forcleaning and inspection), validation of equipment cleaning methods is not expected.

Cleaning validation is expected, as per Section 12.7, once API batches are produced forcommercial use.

Q. Is cleaning qualification sufficient? Would cleaning validation be expected once the APImanufacturing was commercial and not clinical?

Section 12.7 emphasizes that "cleaning procedures should normally be validated." This applies toAPI commercial processes.

Cleaning or any validation is pretty much recognized as being generally inappropriate for APIs forUse in Clinical Trials. Section 19.3 states: "procedures should be in place to ensure thatequipment is calibrated, clean, and suitable for its intended use."

Q. Regarding use of equipment for pre-clinical and clinical, is it allowable to use the sameequipment for both as long as all operations are conducted under Q7A and as long as theworse case scenario is considered for cleaning?


22/123

This could be acceptable as long as you have adequate cleaning procedures to remove residuesof the previous material from the equipment and those procedures have been shown to beeffective.

Q. Is cleaning validation required before the API starting material?

No. Section 1.3 (Scope) of Q7A states "this GMP guidance does not apply to steps prior to theintroduction of the defined API starting material."

Q. What is the minimum adequate interval for monitoring a validated cleaning process?

Section 12.7 emphasizes that "cleaning procedures should be monitored at appropriate intervalsafter validation to ensure that these procedures are effective when used during routineproduction." It is up to the company to define the interval for monitoring.

Q. After calibration and/or maintenance of a process unit, is it required to clean the unitbefore use for processing?

This is not specifically addressed in Q7A. However, using a common sense approach, if thecalibration or maintenance of process equipment exposes the equipment to the environment orintroduces contaminants, it would be prudent to clean the equipment before use.

Q. Vendors of new equipment often ship you a vessel that has been polished. Often theydo not know the chemical constitutes of the polish or of the gases. Our approach tocleaning, in this case, is use of a non-polar solvent to remove organics followed by a polarsolvent to remove inorganics. Is this acceptable?

First of all, we have not discussed this kind of cleaning within the expert-working group. What youhave just described is a perfect example of how to use scientific knowledge, and it sounds like agood approach for dealing with this situation. .

Q. Please elaborate the definition of dedicated equipment. Equipment used on a campaignduring six months can be considered dedicated during this period. Does cleaningvalidation apply if change of product is forecasted?

Dedicated means that it is used solely for one product and only this product is manufactured inthis equipment. In the case of the six-month campaign, it's still a multi-purpose facility. As perSection 5.2, equipment assigned to continuous production or campaign production "should becleaned at appropriate intervals to prevent build-up and carry-over of contaminants (e.g.,

degradants or objectionable levels of microorganisms)." It is up to the company to define theappropriate intervals for cleaning equipment. During this six-month campaign, there may havebeen cleaning between batches, in addition, appropriate cleaning should be conducted after thecampaign is finished to prevent cross-contamination. It would also be appropriate to considervalidation of the cleaning at the end of the campaign.

Section 5.3


23/123

Q. Do you need to calibrate non-critical equipment?

Normally you calibrate the instruments, not equipment. Section 5.3 of Q7A addresses calibrationof critical instruments, but it may be wise to calibrate any instrumentation because measurementsmay have an impact on the overall process. The calibration program may look different for non-critical instrumentation.

Q. Is it the expectation of Q7A that equipment should be calibrated before and after apreventive maintenance is done on a piece of equipment?

There is not necessarily a link between the calibration of instruments and the preventivemaintenance. If the preventive maintenance affects the calibration, then calibration should bedone afterwards.

Q. What types of controls are expected if the calibration of equipment is contracted tooutside agencies?

No different controls than what would be expected in-house. Any contractor providing a GMPservice is an extension of your own company, and there are the same expectations. It's just who'sdoing it, and you still have the responsibility to insure that it's done correctly.

Section 5.4

Q. Regarding computer validation, is it necessary to second check all data into thecomputer system or only the critical data?

Section 5.4 explicitly limits this to critical data and not to all data.

Q. Revalidation of computerized systems. What is required for computerized materialsmanagement systems used to monitor receipts, sampling testing or release of materials?

There is no formal necessity for routine revalidation of computerized systems. The need forrevalidation is based on changes made to the system.

Q. Does Q7A consider 21 CFR Part 11, and what is the regulatory status for compliancewhen inspecting an API manufacturer? Does Q7A specifically address or implycompliance requirements?

Restated, the question is asking whether Part 11 applies to API manufacturers. This questionshould be addressed from both the legality and practicality perspectives. Part 11 is theregulations that set forth the criteria under which the FDA considers electronic records, electronicsignatures and handwritten signatures executed to electronic records to be trustworthy, reliable,and generally equivalent to paper records and handwritten signatures executed on paper. Section11.1(b) states that the regulation applies to (1) records in electronic form that are created,modified, maintained, archived, retrieved, or transmitted, under any records requirements setforth in agency regulations; and (2) electronic records submitted to the agency underrequirements of the Federal Food, Drug and Cosmetic Act and the Public Health Service Act,


24/123

even if such records are not specifically identified in agency regulations (this generally applies toNDA's or ANDA's submitted electronically to the agency)

Herein lies the legality issue. Part 11 cannot be imposed on API manufacturers because in theUnited States we lack a GMP regulation specifically covering the production and control of activepharmaceutical ingredients that sets forth records requirements. Q7A is industry guidance, not a

regulation. It is not legally binding and does not establish requirements. .

However, from a practical standpoint, I would not run to the bank with this. If you are an APImanufacturer with an established paper record and manual signature system and arecontemplating or engaged in converting to an electronic records and electronics signaturessystem, would you not consider it prudent to assure that this new electronic records system istrustworthy, reliable, and generally equivalent to your paper records and handwritten signaturesexecuted on paper? If so, would it not make sense to follow Part 11 as guidance to provide thisassurance?

Section 5.4 of Q7A includes the following expectations for computerized systems:

Computerized systems should be validated Installation and operational qualifications should demonstrate the suitability of hardware

and software

Controls to prevent unauthorized access or changes

Record of any data change made, the previous entry, who made the change and whenthe change was made

Change control for modifications to hardware, software and any other critical components

In addition Q7A gives you some flexibility as it allows for recording of data by a second means inaddition to the computer system. This allows you for example to manually record temperature ona batch record in addition to the instrument recording.

Q. Master batch instructions written by a word processor using something like MicrosoftWord, then printed, reviewed and approved as a hard copy, which is signed and dated.True or false, is this an electronic record?

Q7A (section 5.4) does not define what constitutes an electronic record. But from a Part 11perspective, it depends on is how you are using that document.

Section 6:Documentation and RecordsQ. Section 4.2 requires utility system drawings be available. What level ofdocumentation/drawings is expected? Are piping and instrumentation diagrams sufficient,or are drawings showing actual dimensions expected?

Basic diagrams showing the flow and the tracking of the particular piping throughout the facilityhave been adequate to this point in time. But, identification is the bigger issue. In most cases, theengineering diagrams that the company has obtained from their engineering companies (i.e., thecompanies responsible for designing and constructing the equipment and installing it) wouldsuffice for inspectional purposes. It is important that those engineering diagrams be current andthat they actually show what's in place in the plant at that particular moment. Linking your changecontrol program to updates to your engineering drawings is key to keeping them current.

Q. The CGMP regulations for drug products, 21 CFR 211, specify a written record of major


25/123

equipment cleaning, maintenance, and use shall be included in the individual equipmentlogs. Q7A does not specify individual equipment logs, only records. Does this meanindividual equipment logs are no longer necessary/ expected for API manufacturing?

Section 6.2 states that "records of major equipment use, cleaning, sanitation, and/or sterilizationand maintenance, should show the day, time (if appropriate), product, batch number of each

batch processed in the equipment and the person who performed the cleaning and maintenance.If equipment is dedicated to manufacturing one intermediate or API, then individual equipmentrecords are not necessary. In cases where dedicated equipment is employed, the records ofcleaning, maintenance, and use can be part of the batch record or maintained separately."

It certainly seems to imply that if it is not dedicated, then you don't have those two options. Youonly have the option of having it as part of the batch record.

This is one area where the EWG was attempting to address the uniqueness of APImanufacturing. In many API plants, the entire operation is often controlled from an operation or acontrol room, where you have an electronic board or some type of automatic control system andthe operators sit there and push buttons to control the process. Imagine if, before a particularpiece or reactor was to be used, the operator had to stop the process run down to the second

floor to get the equipment use and cleaning log for that particular reactor. This doesn't soundreasonable. From the GMP perspective, it's important that equipment cleaning and use bedocumented. Where it is documented is up to the company.

Q What microbial controls is the manufacturer required to have in place for an API used inclinical trials for parenteral drug product even if the manufacturer has no responsibility forthe end use of the API?

The same controls that any private company would be expected to have in place since it is anAPI. In today's global market, it would be hard to conceive that an API manufacturer would notknow the intended use of the API provided to a customer. Particularly, in the clinical trial arenathe API manufacturer knows what company is going to receive that API. In many cases, this is

the same corporation, because of your new drug substances, new APIs, both the API and thefinal product are developed by innovative companies. The responsibility to ensure the API isappropriately controlled for use in a parenteral would lie with the Drug Product manufacturer.

Q When making corrections to documented entries, the only requirements are to sign,date, and leave original entry readable. Was there any discussion when writing Q7A toalso record the reason for the change? Will field investigators expect this?

The EWG did discuss it but decided not to include it in Q7A. While it might make sense to recordthe reason, there was concern about how to word the requirement and how it would beinterpreted. (Section 6.1)

Q. Please expand on what is acceptable as true copies. Do you mean actual reproductionsin place of originals, scanned records, or validated computer system accepted by theFDA?

Yes, these all would be true copies; also microfilm, microfiche. Typically with a true copy like that,someone has reviewed the original and the copy and has signed a document somewhere thatindicates that a review has been done and that this is a true copy of the original. (Section 6.1)


26/123

Q Should an Equipment Use Log be maintained for non-dedicated equipment? Should itinclude listing which piece of equipment was used for that particular product at thatparticular moment?

Section 6.2 of Q7A clearly states, "If equipment is dedicated to manufacturing one intermediate orAPI, individual equipment records are not necessary if batches of the intermediate or API follow in

traceable sequence. In cases where dedicated equipment is employed, the records of cleaning,maintenance, and use can be part of the batch record or maintained separately." Thus, it is up toeach company to define the format for maintaining these records.

Q Does the list of raw materials need to include a list of filters and other commodities?Does Q7A require control of durables and consumables?

Q7A does not specifically categorize filters as a raw material. It is important that the batchdocumentation record the specifics of the filters used to facilitate subsequent investigations.(Section 6.3)

Q. Is there a list or reference available that defines objectionable organism?

When we talk about objectionable organisms, we're really thinking of the intended use of theproduct and what organisms would be objectionable based upon that intended use. One exampleis a topical ointment that is placed over an open wound. If you had Pseudomonas contaminationin that ointment, that would be objectionable. So, objectionable really ties into the intended use ofthe drug product and whether it the presence of that organism would be objectionable given thatintended use. USP identifies indicator organisms, pathogens that should be tested for undercertain circumstances.

Q Why was the validation protocol not used as a source document for defining criticalparameters?

Different companies put that documentation in different files. In some companies, that informationis in the development report; in some, it's in the validation protocol; and in some, it's inengineering source documents. We did not try to reference specifically which document wouldidentify, what was critical. It was the EWG 's opinion that the validation protocol was too late fordefining what would be critical. By that point, the development reports should have alreadyidentified from R&D and from engineering what they pre-define as critical areas of concern.(Section 6.4)

Q What degree of validation should be necessary when data is recorded on a computerand a second system, i.e. paper batch record? Differences could exist in the degree ofaccuracy on the computer versus the paper. Also, in deciding if a deviation is critical,would one normally expect to see what requirements are critical versus non-critical pre-defined in the production instructions? If not, what justification would you expect to see?

With the exception of significant figures and timing, there's not going to be much of a difference,normally. Many bulk manufacturers record the data electronically from their distributed controlsystem, and then have critical steps manually recorded by operators for a second check. Inaddition, it makes sure that the operators are focused on these important readings. Certainlycompanies are not prohibited from having two systems operating at the same time.


27/123

Q. Section 6.5 says, " . . . should include complete information relating to production," yetclarification to 6.3 says " . . . short document for recording results;" please explain thediscrepancy.

While it is not specifically mentioned in the Q7A document, some companies have chosen tocontrol their documents in this manner. The white box that was shown on the slide said that

there's no one right way to do it. You can have all of the instructions on one document, and thenyou can have a short document that would have the same numbers, so you just record theresults. That's what is meant by short, as opposed to having blanks throughout the instructionwhere you would record the results, and each completed batch record would be a copy of theinstruction and the results together

A similar situation occurs in drug products regarding the requirements of 211.188(b) where itsays, "documentation that each significant step in the manufacturing, processing, packing, orholding of the batch was accomplished." A good example would be documenting an integrity teston a .22-micron filter. In some companies, in the batch record, it will say perform the integrity testfollowing procedure so-and-so, and the person then goes and looks up the procedure. Othercompanies will detail the actual procedure in the batch records.

Q In deciding if a deviation is critical, would one normally expect to see what requirementsare critical versus non-critical pre-defined in the production instructions? If not, whatjustification would one expect to see?

Remember the definition for critical. If it were something that affects the quality of the API, youwould probably want to include it in the batch record.

Q. Is there a distinction between a significant and a critical step? Q7A does not define"significant."

Q7A does define critical, and what you will probably find as you're doing your process validation

or going through your development reports is that there are a lot of variables with different levelsthat they will contribute. What EWG tried to address in Documentation and Records, is the criticalsteps. For your own process consistency, you will probably keep records of steps in addition tothe critical steps. But, we have not stated that there is a second signature required, for example;the level of documentation is not the same. You will probably have subsets, such as keyparameters, or significant parameters, or whatever.

Q If a data entry were missed, not done directly after performing the activity, is noted andcorrected subsequently, i.e. the next day, would you expect a deviation report, in addition?

It is important to know what kind of data we're talking about, but you've probably got to havesomeone in Quality look at what you're doing (i.e., what it is you're putting in later). So, first, whatit is; and second, how do you know that's the correct number the next day? Typically, that wouldbe written up in the format of a deviation, because then you could have all the signatures on theappropriate page.

Q. What are the expectations for second-person review of data entry in the laboratoryincluding weighing, adding, etc.?


28/123

Section 6.6 of Q7A indicates that the review of laboratory records should include "the date andsignature of a second person showing that the original records have been reviewed for accuracy,completeness, and compliance with established standards." Typically, second-signature onweighing in the laboratory is not done for GMP reasons; if you're handling narcotics or controlledsubstances, that's a totally different issue. In order to review accuracy, there typically has to be acheck on the calculations. Review for completeness would be to confirm that all of the work hasbeen done, and everything has been filled out fully. Compliance with established standards wouldbe accomplished through comparison of the result against the expected result.

Having a second person always watching over your shoulder to witness everything you do is anunreasonable expectation. This could create havoc in the laboratory and probably impede theproper and efficient analysis of materials and drug products in the lab.

Q What is the difference between official and established specifications? Are these termsinterchangeable?.

That is an example of putting in wording to allow the EWG to reach consensus. "Official" wasmeant to refer to the specifications that would be in the Pharmacopoeia, and "established" being

the ones established by the company. "Official" could also mean specifications that are submittedin a regulatory submission to the regulatory bodies.

Q. (Section 6.4) To what level of detail should the manufacturing master productioninstructions be written with the assumption that the operator has a Ph.D. or is totallyignorant?

The instructions are written not based on intelligence level, but are written with a level of detailsuch that the operation can be carried out and that in reviewing at a later time, you can knowexactly what the history of that batch was. Some of the detail may be included in area SOP'srather than being specified in the master production instruction.

Q. A batch record should be checked before it is issued to production. Can this check bedone by production or must it be done by quality?

Q7A (section 6.5) says that the quality unit must check the master production instruction, but forthe batch production record, it is silent on who much do the check. Therefore, either quality orproduction could do this.

Q. Does batch record generation software have to be Part 11 compliant? Can Microsoftproducts be used to generate these batch records? Do batch record generation programshave to be validated?

Q7A-Section 6.5 says that the batch production record should be checked that it is an accuratereproduction. (Note: See section 5.4 for other questions related to Part 11 compliance.)

Q. Should raw materials have traceability so as to establish the raw material history, etc.?

Yes. Section 6.3 recommends that you show where the raw material was used.


29/123

Q. Would you expect to see an official document from the API manufacturer describing thestarting material and its position in the synthetic pathway?

Yes. Section 1.3 states that the company should designate and document th

q7 starting material questions and answers

Documents